Article pubs.acs.org/est
Short-Chain Chlorinated Paraffins in Zurich, SwitzerlandAtmospheric Concentrations and Emissions Pascal S. Diefenbacher,†,‡ Christian Bogdal,*,†,§ Andreas C. Gerecke,‡ Juliane Glüge,† Peter Schmid,‡ Martin Scheringer,†,∥ and Konrad Hungerbühler† †
Institute for Chemical and Bioengineering, ETH Zurich, Vladimir-Prelog-Weg 1, CH-8093 Zurich, Switzerland Empa, Swiss Federal Laboratories for Materials Science and Technology, Ü berlandstrasse 129, CH-8600 Dübendorf, Switzerland § Agroscope, Institute for Sustainability Sciences ISS, CH-8046 Zurich, Switzerland ∥ Environmental Chemistry and Substance Dynamics, Leuphana University Lüneburg, 21335 Lüneburg, Germany ‡
S Supporting Information *
ABSTRACT: Short-chain chlorinated paraffins (SCCPs) are of concern due to their potential for adverse health effects, bioaccumulation, persistence, and long-range transport. Data on concentrations of SCCPs in urban areas and underlying emissions are still scarce. In this study, we investigated the levels and spatial distribution of SCCPs in air, based on two separate, spatially resolved sampling campaigns in the city of Zurich, Switzerland. SCCP concentrations in air ranged from 1.8 to 17 ng·m−3 (spring 2011) and 1.1 to 42 ng·m−3 (spring 2013) with medians of 4.3 and 2.7 ng·m−3, respectively. Both data sets show that atmospheric SCCP levels in Zurich can vary substantially and may be influenced by a number of localized sources within this urban area. Additionally, continuous measurements of atmospheric concentrations performed at one representative sampling site in the city center from 2011 to 2013 showed strong seasonal variations with high SCCP concentrations in summer and lower levels in winter. A long-term dynamic multimedia environmental fate model was parametrized to simulate the seasonal trends of SCCP concentrations in air and to back-calculate urban emissions. Resulting annual SCCP emissions in the city of Zurich accounted for 218−321 kg, which indicates that large SCCP stocks are present in urban areas of industrialized countries.
■
Convention on Persistent Organic Pollutants.5 Because CPs are produced, used, and emitted into the environment as complex mixtures containing thousands of isomers, enantiomers, and diastereomers, detection and quantification at trace level in environmental samples is extremely difficult.7 Compared to other chlorinated organic compounds, information on environmental levels and fate of CPs is scarce.2,8,9 Presently, there exist only a few detailed studies that report atmospheric SCCP concentrations in Europe,10,11 North America,12,13 Asia,14−17 the Arctic,18,19 and Antarctic.20 The understanding of existing sources and the atmospheric fate of SCCPs is still incomplete. Here, we measured SCCPs in ambient air in the city of Zurich, Switzerland and employed a mass-balance model to understand the environmental distribution of SCCPs and to back-calculate their emissions. The city of Zurich serves as a case study, as it is representative of urban
INTRODUCTION Chlorinated paraffins (CPs) are synthetic mixtures of chlorinated n-alkanes with a chlorination degree varying between 30 and 70% by weight.1 According to the carbon chain length, they are subdivided into short-chain (C10−C13, SCCPs), medium-chain (C14−C17, MCCPs), and long-chain (C18−C30, LCCPs) CPs. CPs are mainly used as high pressure additives in metal working applications and as a secondary plasticizer in plastics (e.g., PVC, polyester, neoprene), followed by their application as adhesives and sealants in textiles and polymeric materials, as well as in leather production and paints. The global production of CPs was approximately 300 000 tonnes in 1985.2 By 2009, China was the largest producer with about 1 000 000 tonnes per year.3 In Switzerland, a widespread use was in joint sealants (e.g., splices of buildings), where CPs were often used as substitutes for the banned polychlorinated biphenyls (PCBs).1 Among CPs, SCCPs show the highest toxicity and strongest bioaccumulation.4,5 Further, SCCPs are persistent in the environment and have the potential to reach remote areas through long-range atmospheric transport.1,6 SCCPs are proposed for listing under the Stockholm © XXXX American Chemical Society
Received: April 29, 2015 Revised: July 9, 2015 Accepted: July 16, 2015
A
DOI: 10.1021/acs.est.5b02153 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
°C in sealed plastic bags. Before Soxhlet extraction with DCM/ Hex (1/1), 5 ng of 13C10-labeled 1,5,5,6,6,10-hexachlorodecane (CIL Cambridge Isotope Laboratories, Andover, USA) was added as internal standard. The extracts were cleaned by a multilayer-column packed with anhydrous Na2SO4 (2 g), activated silica gel (2.8 g), and acidic silica gel (1.7 g, 44% concentrated sulfuric acid) from top to bottom. Subsequently, the extracts were reduced to 1 mL using a rotary evaporator. To separate SCCPs from other organochlorine compounds that could influence analysis, a second column containing Florisil (3 g, deactivated with 1.5% H2O) topped with anhydrous Na2SO4 (2 g) was applied. The first fraction was eluted with 15 mL nhexane and analyzed for PCBs as reported elsewhere,23 whereas the second fraction containing CPs as target compounds was eluted with 15 mL dichloromethane. The collected extracts were solvent exchanged to toluene and concentrated to approximately 20 μL. Prior to analysis, 2 ng of ε-HCH (Dr. Ehrenstorfer, Augsburg, Germany) was added as recovery standard. Instrumental Analysis and Quantification. SCCP concentrations were determined by gas chromatography coupled to electron capture negative ionization high-resolution mass spectrometry (GC/ECNI-HRMS) using a MAT 95 HRMS (Thermo Finnigan MAT, Bremen, Germany). The separation was conducted in a self-manufactured glass capillary column (20 m × 0.28 mm) coated with a 0.15 μm thick film of 15% diphenyl-/ 85% dimethyl-polysiloxane (PS 086, Fluka, Buchs, Switzerland) with hydrogen as carrier gas (40 kPa) and argon as the reagent gas. The temperature program started at 110 °C, held for 1 min and ramped to 310 °C at 10 °C·min−1, and held for 3 min. Injector temperature was set to 260 °C and the ion source temperature was set to 120 °C. An aliquot of 2 μL was injected in splitless mode. SCCP mixtures with a chlorine content of 51.5%, 55.5%, and 63% (Dr. Ehrenstorfer, Augsburg, Germany) were used as external calibration standards. Identification and quantification of SCCPs was performed by detecting isotope clusters of the [M−Cl]− ions in the selected ion monitoring (SIM) mode, according to the method of Tomy et al.26 However, we slightly modified the selection of quantitation ions, as for three of these ions, we obtained better shapes for the second most abundant signal and used the latter one for quantification (SI Table S2). Afterward, the signal of the most frequent congener group in a sample was quantified relative to the signal of an external standard mixture to determine the total SCCP content of the sample. This method requires a standard that resembles the sample as much as possible. The availability of commercial SCCP standards is very limited, thus it is not always possible to obtain a suitable standard. Unfortunately, application of a standard with a slightly different chlorination degree can lead to significant deviations in resulting amounts, because the instrumental response is highly dependent on the chlorination degree of the analytes.7 Therefore, we additionally corrected the results by a procedure similar to the methodology of Reth et al.27 This procedure makes it possible to compensate for differences in the chlorination degree of the environmental samples and reference mixtures. Using six SCCP standard mixtures with a chlorination degree of 51.5−63% (w/w), a linear correlation between the relative response factor and the chlorination degree of standard mixtures was established (SI Figure S2). According to the respective chlorination degree, a relative response factor was determined for every sample and standard
areas in industrialized countries with appreciable usage of chemical additives such as CPs. The mass-balance model applied has previously been used and validated for organohalogen contaminants in ambient air in Zurich.21−24 To better understand the sources and fate of SCCP in urban air, we investigated the spatial distribution of concentrations in air within the city of Zurich and recorded seasonal trends over a three-year period. This study is the first to report temporally resolved SCCP emission fluxes of an urban area.
■
MATERIALS AND METHODS Passive Air Sampling. The spatial distribution of SCCP concentrations in air was investigated with two sampling campaigns using polyurethane foam (PUF) based passive air samplers (PAS) deployed within the city of Zurich. A detailed description of these PAS can be found in Diefenbacher et al.23 The first sampling campaign for SCCPs was performed in spring 2011 (March to May) with PAS deployed at 23 locations. Thereof, seven sites were located in the outskirts of the city (A1−A7), another seven were randomly distributed in the urban area (B1−B7). The remaining nine sites were located in a specific neighborhood (C1−C9) to account for smallerscale differences in atmospheric SCCP levels (Supporting Information, SI, Figure S1). The samplers were either placed at a height of 1.5−3 m above ground or attached to a building for a time period of 29 to 42 days. All PAS were installed in a manner that air could circulate freely around them. A second sampling campaign was performed in spring 2013 (March to May) at the same locations with a deployment time of 40 to 49 days. A detailed description of the individual sampling sites and exposure dates is provided in SI Table S1. In addition, measurements were conducted continuously from March 2011 to December 2013 at one site (Kaserne, 47.377°N, 8.530°E) in the center of Zurich with relative sampling intervals of approximately 6 weeks. For these measurements, PAS were situated in a large courtyard of a public building. This site was previously used for ambient air sampling as described by Diefenbacher et al.23 A sampling rate (Rs) of 4.2 m3·d−1 was used to calculate SCCP concentrations in air. This Rs value is based on extensive PUF calibration experiments by Li et al.,14 who employed the same type of samplers. To check the applicability of this Rs value for our setup, control measurements were conducted with an active high-volume air sampler deployed next to the PAS. The resulting Rs value was 3.9 m3·d−1, indicating that the literaturederived Rs value is accurate for this study. Further, active air sampling allowed us to investigate gas-particle partitioning of SCCPs. Depending on the average air temperature during the sampling period, the fraction of total SCCPs in the particle phase was between 12 and 40%. It is known, that the particlephase fraction is significantly higher during wintertime, while SCCP occurs mostly in the gas-phase during summer.16,20 Recently, Markovic et al.25 evaluated the particle infiltration efficiencies of the PAS used in this study and observed that they are able to acquire a representative sample of ambient particles. This finding implies that PAS are suitable for measuring chemicals that are partly particle-associated. Thus, atmospheric SCCP concentrations shown in this work always include gasand particle-phase fractions. Sample Preparation and Processing. Prior to application, PUF disks were precleaned by Soxhlet extraction with dichloromethane/n-hexane (DCM/Hex) (1/1). After deployment, PUFs were wrapped in aluminum foil and stored at −20 B
DOI: 10.1021/acs.est.5b02153 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
Figure 1. Box-and-whisker plot of the SCCP concentrations in air measured in 2011 (a) and 2013 (b). The box stretches from the 25th-percentile to the 75th-percentile and contains the median in between. The whiskers end at the minimum and maximum observed values or at most 1.5 times the interquartile range. Outliers lower or higher than 1.5 times the interquartile range are represented separately by the dots. (c) Gray bars represent the measurements performed in spring 2011, black bars show the results of the campaign conducted in 2013 at identical sampling sites. A detailed description of the sampling sites including a map is provided in the SI.
mixture, and the resulting amounts of SCCPs were corrected for differences in the chlorination degree. More details and equations are provided in the SI. Quality Assurance. To minimize the risk of contamination all glassware used in this study was heated to 450 °C. Field blanks were derived by transporting pre-extracted PUF disks to the sampling sites and exposing them to ambient air for a few minutes. Afterward, these field blanks were extracted and analyzed in the same way as the samples. The method detection limit (MDL) was defined as the mean level of these field blanks plus 3 times the standard deviation. Assuming a deployment time of 40 days and sampling rate of 4.2 m3·d−1 the MDL value for SCCPs in air was estimated to be 0.5 ng·m−3. As the mean concentrations of the blanks were much lower than the SCCP concentrations in ambient air, results were not blank corrected. The recoveries of 13C10-labeled 1,5,5,6,6,10-hexachlorodecane from the atmospheric samples were in the range of 50−120%. Model Design and Parametrization. The multimedia mass-balance model used in this study tracks partitioning and fate of SCCPs in the region of Zurich, Switzerland (400 000 inhabitants on 100 km2). In an earlier study by Diefenbacher et al.,23 this model was presented in detail; it was successfully used to reproduce PCB concentrations in air over a time period of three years and to derive the atmospheric emissions. The model is composed of environmental compartments that include atmosphere, soil, water, sediment, and vegetation. The atmosphere is further subdivided into lower air, upper air, and free atmosphere (SI Figure S4). Because the purpose of the model is to determine SCCP concentrations in urban ambient air, the focus of this work lies on the lower air compartment. To allow a comparison with the measured SCCP levels, model results represent concentrations in the bulk air phase, hence including gas- and particle-phase SCCPs. The main environmental processes are wind-driven advective inflow and outflow of airborne chemical, bidirectional diffusive exchange of
chemical between compartments, settling of particle-bound chemical in the air and water compartments, and degradation of chemical in all compartments. The purpose of this model is to calculate highly resolved SCCP concentrations over a time period of three years (2011− 2013). A feature of this model is the temporal variability of the three atmospheric layers. The height of the nocturnal stable boundary layer (SBL) and the convective boundary layer (CBL) that is formed during the day were derived from weather prediction data and were updated for every day (SI Figure S5). Further, air temperature and wind speed were determined by averaging hourly measurements from four nearby meteorological stations. As transformation rate constants and the physicochemical properties of the SCCPs depend on temperature, they were continuously adapted to the actual environmental temperature used in the model. The mean annual SCCP concentration in background air was determined at a sampling station in the rural surroundings of Zurich (Tänikon, 30 km distant from the city center). Seasonal changes of the SCCP concentration in background air were implemented with a sinusoidal forcing function (see SI). The overall SCCP emissions in the city were modeled as temperature-dependent volatilization flux from a hypothetical pool of pure-phase chemicals. The surface area of this pool was the only input parameter in the models that was not predefined, as it was adjusted to obtain model results for the absolute SCCP concentrations in air that are as close as possible to the measurements. Changes in the pool’s surface area directly affect SCCP concentrations in air but do not influence the seasonal trends, which result from the independent and predefined meteorological parametrization of the model. The surface area of the emission pool was adjusted to minimize the total deviations of measured and model air concentrations, using the method of least-squares minimization. Therefore, the surface area of this pool is specific for each SCCP congener group and C
DOI: 10.1021/acs.est.5b02153 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
Figure 2. Average concentrations of SCCPs in air (gray bars) and average air temperatures (black line) measured at site B1. Red dots stand for duplicate measurements. Dates on the x-axis represent starting points of the sampling periods with a respective duration of 6 weeks.
the urban median. Only at sampling site B7 the measured SCCP concentrations in air were significantly higher than the median in both years. Here, the PAS was placed near a hydroelectric power station, and the increased concentrations in air are possibly reflecting past usage of SCCP. Atmospheric levels at site B4, however, were below the median in 2011 and exhibited the highest measured level in 2013. This may be caused by the demolition of a building next to the sampling site in 2013. The other sampling sites show SCCP levels that are within a factor of 2 of the median in both years or at least in one year and no strong SCCPs sources in their surrounding are likely. Overall, median SCCP concentrations in 2011 were a factor of 1.6 higher than in 2013, which may be explained by differences in meteorological conditions. Air temperatures were slightly higher in 2011, thus increasing the volatilization of SCCPs. Additionally, the average daily rainfall in spring 2011 was 1.0 mm, whereas in spring 2013, it accounted for 4.6 mm. It is known that rainfall can induce a significant wash-out of compounds associated with atmospheric particles.31 Another factor possibly influencing SCCP concentrations in air is the wind direction that was not determined at individual sampling sites. It seems that SCCP emissions in the city of Zurich are caused by a number of diffusive sources that are distributed over the city. This result is in agreement with the presumption that the most widespread SCCP use in Switzerland was in joint sealants of buildings, and industrial application with high SCCP emissions are scarce.1 An important observation is that SCCP levels measured at sampling site B1 were very close to the median in both sampling campaigns. Therefore, this site is deemed a good representative for the overall urban air concentration and was further used to study seasonal variations of air concentrations. Seasonal Variation. SCCP concentrations in ambient air measured at site B1 ranged from 0.78 ng·m−3 in January 2013 to 5.7 ng·m−3 in April 2011 (Figure 2). SCCP levels in air showed a seasonal trend with average concentrations of 3.8 ± 1.9 ng·m−3 during the summer half-year (April−September), while lower concentrations of 1.8 ± 1.5 ng·m−3 were found during the winter half-year (October−March). The mean annual concentration in air over the complete sampling period of three years was 2.9 ng·m−3 (median: 3.1 ng·m−3). The
was kept constant for the overall modeling time of three years. Further details on the parametrization of the model can be found in Diefenbacher et al.,23 furthermore a list of the input parameters specific to this study is presented in the SI. The influence of uncertainty in input parameters on the model results was assessed by an uncertainty analysis according to MacLeod et al.28 and a 95% confidence interval was calculated for model outputs. Confidence factors for each input parameter are given in SI Table S3 and S4. SCCP Congener Groups Selected for Modeling. Under the assumption that SCCPs do not include dichlorinated carbon atoms and branched chains, SCCPs are a composite of 7820 isomers and 46 congener groups.26,29 Owing to the complexity of SCCP mixtures, partition coefficients and degradation half-lives vary considerably among different SCCP congeners. Our long-term model uses a selection of four congener groups with well-defined physicochemical properties that cover a representative fraction of the total atmospheric SCCP composition (SI Figure S6). Thus, the most abundant congener groups of different carbon chain lengths (C10H16Cl6, C11H17Cl7, and C12H19Cl7) were selected for modeling. Further, we include the most abundant congener group containing eight chlorine atoms (C11H16Cl8) to cover a broader range of chlorination degrees (60.3−65.7%). Partition coefficients of each congener group were derived from Glüge et al.30 and are represented by the geometric mean of the isomers considered in each group. All physicochemical properties are specified in SI Table S3.
■
RESULTS AND DISCUSSION Spatial Distribution within the City. SCCP concentrations in air measured in spring 2011 ranged from 1.8 to 17 ng·m−3, with a median concentration of 4.3 ng·m−3. In spring 2013, the median concentration was 2.7 ng·m−3 and values ranged from 1.1 to 42 ng·m−3 (Figure 1). Average air temperatures of 11.5 and 9.6 °C and average wind speeds of 1.4 and 1.5 m·s−1 were measured in 2011 and 2013, respectively. Both data sets show that concentrations at single spots within an urban area can vary substantially, as the maximum levels were by factors of 9 (spring 2011) or even 37 (spring 2013) higher than minimum levels. Nevertheless, most of the measurements deviated by a factor of less than two from D
DOI: 10.1021/acs.est.5b02153 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology Table 1. Summary of Atmospheric SCCP Concentrations in Switzerland and Other Regions surrounding (n)a
region and year b
a
Switzerland, Zurich (2011) Switzerland, Zurich (2013)b Switzerland, Zurich (2011−2013)b Switzerland, Tänikon (2013)b UK, Lancaster (1997)11 UK, Lancaster (2003)10 China, Beijing (2011)16
urban (23) urban (23) urban (1) rural (1) semirural (1) semirural (1) urban (1)
China (2008)14 India (2006)15 Pakistan (2009)15 Japan (2008)14 South Korea (2008)14 Antarctica (2013)20
urban (16), rural (2) urban (13), rural (7) urban (9), rural (1) urban (10), rural (2) urban (5), rural (2) remote (1)
sampling season spring spring all seasons spring all seasons spring winter summer spring and fall winter winter spring and fall spring and fall winter
mean (range) in ng·m−3 5.8 6.4 2.9 0.67 0.32 1.1 7.7 200 137 10 5.1 2.3 2.1 0.01
(1.8−17) (1.1−42) (0.78−5.7) (
